I appreciate your passion, but you have a lot of things wrong. Chiefly this: HTML5 is not a singular entity, but rather a broad description of several emerging web technologies. I launched one of the first commercial HTML5 sites on the Web, and in the years since, the meaning of "HTML5" has already changed significantly, and that's okay.
The semantic HTML tags are less than useless, because they're based on a now obsolete statistical analysis of common ids/classes.
Nearly every site on the Internet has a header and footer, though; sides/sidebars are pretty common too.
Effectively you can't infer any meaning from the semantic tags anymore because it's such a fucking hash up as to what they're even going to mean due to different people going different ways on them.
I'd like to introduce you to the complexity of human language. This isn't a problem specific to HTML5, or HTML, or heavens, even markup languages. People use words and descriptions in different ways. There is not now, never was, and never will be a singular and coherent semantic structure. What's an article to one person is a section to another.
It's important to note that the specs you rail against go to great lengths to avoid enforcing a rigid semantic context. The new HTLM5 elements aren't complete, and nobody ever claimed they were. Their role is purposefully vague, giving web authors new markup tags that infer some semantic structure without enforcing precision. I mean, how many hundreds of roles does the good old tag play? Semantically it indicates a heading of some sort, but what that means, precisely, is intentionally undefined -- and it works great.
The solution to your problem is XML, wherein authors can markup every document with precisely-semantic tags. We tried that with XHTML, and it turned out to be a solution in search of a problem. HTML, with its imprecision, is fine for markup because it's good enough and broad enough that we can do what we want. And when it isn't -- for example, when it fails to define a drawing canvas for complex visual manipulation -- then a new branch (HTML5) was developed. Repeat ad infinitum.
Also, note that I'm talking about just the HTML5 spec here, I have little problem with for example, CSS3.
This is where your argument runs off the rails. You're incensed by HTML5, which has a pretty widely-accepted core implementation, but you have "little problem" with CSS3, which is also incomplete and an ever-changing amalgam of over 50(!) modules, of which only four are in a somewhat-settled state of definition? Come on.
Imagine my horror when I encountered the term "Responsive web design" the other day, apparently it's about creating device/resolution independent web designs. Yeah, great one kiddies who coined that term, we already figured this out some years back.
The "kiddie" who developed responsive web design was Ethan Marcotte, a highly-respected designer and one of the principle authors of new web design techniques. Responsive design was first presented in A List Apart, the leading journal for web design techniques. Responsive web design chiefly involves the use of CSS3 (remember? that thing you have "little problem" with?) media queries in its implementation. And it has nothing to do with resolution independence, but rather, about dynamically adapting both content and layout to a variety of different usage roles (desktop, mobile, etc.).
Again, I appreciate your passion, but it's heavily misinformed. Sure, HTML5 has problems, and they're no secret -- they've been debated ad nauseum since WHATWG first convened. But the fact is that there is a generally-agreed-upon HTML5 baseline, it's in wide use across the Web, and it's enabled some pretty terrific new sites and apps. This will continue, regardless of hand-wringing about standards and forks. The Web adapts.